q2: Q2 predictive relevance statistics

In matrixpls: Matrix-Based Partial Least Squares Estimation

Description

Calculates Q2 predictive relevance statistics based on comparing predictions and real data.

Usage

q2(originalData, predictedData, model = NULL)

Arguments

originalData	A matrix or a data.frame containing the original data.
predictedData	A matrix or a data.frame containing the predicted data that are compared against the original data to calculate the predictive relevance statistic.
model	There are two options for this argument: 1. lavaan script or lavaan parameter table, or 2. a list containing three matrices inner, reflective, and formative defining the free regression paths in the model.

Details

The Q2 statistic is calculated as 1-sse/sso where sse is the sum of squared prediction errors based on comparison of the originalData and predictedData and sso is based on prediction with mean. If the predicted data contain the groups attribute, which indicates the groups used in blindfolding or cross-validation, the means are calculated separately for each group excluding the predicted group from the calculation.

Value

A list with total, block, and indicator elements containing the Q2 predictive relevance statistics for the full dataset, for each indicator block, and for each indicator

matrixpls documentation built on April 28, 2021, 5:07 p.m.